Hi,
I can just confirm what Charles explained. I can just add that one of the reasons for removing the protocol messages was also to reduce confusion, since their only use would have been for drawing sequence diagrams (for which I see a rather limited use of). Having them in the protocols would mainly confuse inexperienced modelers. So left are the only protocol messages that really matters, i.e. to define triggers in state machine based rtBound, rtUnbound and timeout.
Regarding your question if the RTBaseCommProtocol will remain in the Internal package, I get a feeling that you ask this related to how to identify and locate it. May I remind you that I wrote a Bugzilla to track the decision of finding a well defined principle for identifying the system protocols and the base protocol. See
https://bugs.eclipse.org/bugs/show_bug.cgi?id=477721
Please also note that this Bugzilla for tracking this decision blocks a number of other Bugzillas which all should rely on this. I actually realised that I had not made it block the Bugzilla related to the combined trigger dialog, so I just did that to make it clear this we really need to settle this. We had a mail discussion regarding this, but we never came to a conclusion, i.e. whether we should have a principle based on convention, e.g. how the library is packaged, or some more explicit principle, e.g. based on stereotypes (or possibly keywords to avoid having to maintain a separate profile).
/Peter Cigéhn